Approximation and regular perturbation of optimal control problems via Hamilton-Jacobi theory (Q1179159)

This paper is based on some new results on the convergence of the viscosity solutions of the Hamilton-Jacobi equation \(u_ h+H_ h(x,u_ h,Du_ h) = 0\) to the solutions of \(u+H(x,u,Du) = 0\) when \(h \to 0\). In particular, under suitable assumptions the authors give an estimation of the convergence rate \(\sup | u(x) - u_ h(x)| \leq Ch^ \beta\) for some \(\beta > 0\). These results are proved to be useful in various problems arising in control theory. Their applications cover in particular infinite horizon control, problems with exit times, two- players zero-sum games, pursuit and evasion problems and time optimal problems. In particular, the authors apply their results to the study of certain regularization procedures for the minimal time function in the case of linear processes. More generally, they consider the effect on the minimal time function when the system equations are subject to small perturbations.